Molecular Ecology (2014) 23, 1153–1166

doi: 10.1111/mec.12664

The ancient Britons: groundwater fauna survived extreme climate change over tens of millions of years across NW Europe C A I T R
I O N A E . M C I N E R N E Y , * † ‡ L O U I S E M A U R I C E , § A N N E L . R O B E R T S O N , ¶ € RG ARNSCHEIDT,† CHRIS VENDITTI,*†† JAMES S. G. DOOLEY,‡ LEE R. F. D. KNIGHT,** JO THOMAS MATHERS,* SEVERINE MATTHIJS,‡‡§§ KARIN ERIKSSON,† GRAHAM S. € NFLING* P R O U D L O V E ¶ ¶ and B E R N D H A *Department of Biological Sciences, Evolutional Biology Group, The University of Hull, Hull, HU6 7RX, UK, †Environmental Sciences Research Institute, University of Ulster, Cromore Road, Coleraine, BT52 1SA, Northern Ireland, àBiomedical Sciences Research Institute, University of Ulster, Cromore Road, Coleraine, BT52 1SA, Northern Ireland, §British Geological Survey, Maclean Building, Crowmarsh Gifford, Wallingford OX10 8BB, UK, ¶Department of Life Sciences, Roehampton University, Holybourne Avenue, London SW15 4JD, UK, **No. 1, The Linhay, North Kenwood Farm, Oxton, Near Kenton, Devon EX6 8EX, UK, ††School of Biological Sciences, University of Reading, Reading RG6 6BX, UK, ‡‡Royal Belgian Institute of Natural Sciences, Vautierstraat 29, Brussels 1000, Belgium, §§Amphibian Evolution Lab, Department of Biology, Vrije Universiteit Brussel, Pleinlaan 2, Brussels 1050, Belgium, ¶¶Department of Entomology, The Manchester Museum, University of Manchester, Manchester M13 9PL, UK

Abstract Global climate changes during the Cenozoic (65.5–0 Ma) caused major biological range shifts and extinctions. In northern Europe, for example, a pattern of few endemics and the dominance of wide-ranging species is thought to have been determined by the Pleistocene (2.59–0.01 Ma) glaciations. This study, in contrast, reveals an ancient subsurface fauna endemic to Britain and Ireland. Using a Bayesian phylogenetic approach, we found that two species of stygobitic invertebrates (genus Niphargus) have not only survived the entire Pleistocene in refugia but have persisted for at least 19.5 million years. Other Niphargus species form distinct cryptic taxa that diverged from their nearest continental relative between 5.6 and 1.0 Ma. The study also reveals an unusual biogeographical pattern in the Niphargus genus. It originated in north-west Europe approximately 87 Ma and underwent a gradual range expansion. Phylogenetic diversity and species age are highest in north-west Europe, suggesting resilience to extreme climate change and strongly contrasting the patterns seen in surface fauna. However, species diversity is highest in south-east Europe, indicating that once the genus spread to these areas (approximately 25 Ma), geomorphological and climatic conditions enabled much higher diversification. Our study highlights that groundwater ecosystems provide an important contribution to biodiversity and offers insight into the interactions between biological and climatic processes. Keywords: ancestral state reconstruction, Bayesian dating analysis, cave, phylogeography, subterranean Received 18 October 2013; revision received 23 December 2013; accepted 30 December 2013

Introduction Global climate has changed significantly throughout the Cenozoic (65.5–0 Ma) with glacial cycles during the Correspondence: Bernd H€anfling, Fax: +44-1482-465458; E-mail: [email protected] © 2014 John Wiley & Sons Ltd

Miocene, Pliocene and Pleistocene (Zachos et al. 2001a, 2008; Louwye et al. 2008). Precipitation also fluctuated from extended arid (13.2–11.5 Ma) to very wet conditions (10.2–9.8 Ma; B€ ohme et al. 2008). Fauna, for example ectothermic vertebrates and freshwater crustacea experienced major range shifts or extinctions, and ecosystems were dramatically modified (B€ ohme 2003; Klaus & Gross

1154 C . E . M C I N E R N E Y E T A L . 2010). During the Pleistocene glaciations (2.59–0.01 Ma), large areas of the northern hemisphere were covered by glaciers or permafrost and were uninhabitable, with particularly marked biogeographical impact in northern Europe (reviewed in Provan & Bennett 2008). Britain and Ireland are a prime example illustrating this ecological impact, with repeated covering by glaciers and permafrost greatly limiting the persistence of terrestrial species. These islands are likely to have been isolated during interglacials, at least since the formation of the English Channel approximately 0.45 Ma (Gupta et al. 2007), preventing dispersal of terrestrial fauna from the continent. Strong palaeontological and genetic evidence indicates that the majority of the current fauna of Britain and Ireland arrived from mainland Europe following the Pleistocene glaciations, dispersing across a land bridge with continental Europe during the short period after ice retreat and before the bridge was submerged by rising sea levels (Yalden 1982; Hewitt 2004). Consequently, Britain and Ireland have always been thought to have limited endemic biodiversity. However, the biodiversity of groundwater ecosystems may challenge this orthodoxy, with evidence from North America (Holsinger et al. 1983) and Iceland (Kornobis et al. 2010), suggesting that groundwater ecosystems may occur under glaciated areas. Moreover, species are present in formerly glaciated areas, indicating that they must either have survived in refugia or dispersed there since glaciations (Galassi et al. 2009; Martin et al. 2009). As with recent advances in our understanding of deep ocean vent ecosystems (Van Dover et al. 2002; L
opezGarcia et al. 2003; Dubilier et al. 2008), groundwater ecosystems may offer novel insights into fundamental ecological and evolutionary processes. In this study, we use a Bayesian phylogenetic approach, which shows that groundwater fauna must have persisted through glacial periods in Britain and Ireland within refugia. Furthermore, we show how groundwater ecosystems may have developed across Europe in response to changing climatic and geomorphological conditions. Finally, we demonstrate that the biogeographical pattern of diversity across Europe is unexpected, with increasing phylogenetic diversity at higher latitudes. Our study focuses on amphipod crustacea, which are a major component of subterranean ecosystems and offer a tractable model for investigating ecological and evolutionary processes within this challenging environment. The largest genus among them is Niphargus (Amphipoda: Niphargidae) with over 300 described species distributed across most of Europe (V€ain€ ol€a et al. 2008). Niphargus are stygobites (obligate groundwater inhabitants), approximately 0.3–3.0 cm in length, which are adapted to live in subterranean environments. They are blind, lack pigmentation and have elongated appendages

(Fig. 2a). Previous phylogeographical studies of Niphargus have demonstrated high levels of endemism and cryptic diversity at small geographical scales (Fiser et al. 2008; Trontelj et al. 2009), suggesting limited dispersal and long-term persistence of local populations, as well as morphological convergence for adaptations to the subterranean environment (Trontelj et al. 2012). Only six taxa of Niphargus are currently known from Britain and Ireland (Robertson et al. 2009). Here, we show that two species endemic to Britain and Ireland (N. glenniei and N. irlandicus) are far older than previously thought, suggesting persistence through extreme climatic and geomorphological changes over at least 19 million years. Furthermore, those species thought to have been wideranging European species (N. aquilex, N. fontanus, N. kochianus) are in fact also ancient British endemics.

Material and methods Sampling A modified Cvetkov net sampler, notenboom sampler or baited traps were used to collect samples from boreholes, springs and wells. A total of 454 Niphargus specimens were preserved in >70% ethanol (Fig. 1, Table S1, Supporting information), comprising samples from 63 populations (222 individuals) in Britain and Ireland including five of the six species present. We were unable to obtain sufficient samples for DNA extraction of the rare N. wexfordensis. Additionally, 224 individuals from 47 populations and five species were collected from Belgium, the Netherlands, Germany and France (Fig. 1, Table S1, Supporting information) including all species known to co-occur in Britain and continental Europe (N. aquilex, N. fontanus, N. kochianus). Furthermore, samples were obtained from two species that occur in the vicinity of Britain located in France, but for which no DNA sequence data existed (N. pachypus, one population, two individuals; N. forelli, two populations, four individuals). Samples from published data sets covered largely the central and south-eastern part of the distribution and included data of 185 populations from 74 described species (Fig. 1, Table S1, Supporting information).

De novo sequencing and data sets for phylogenetic analysis Genetic variation of Niphargus was assessed at two mitochondrial genes, cytochrome oxidase subunit I (COI) and 16s rRNA (16S) and the nuclear small subunit 28s rRNA (28S; for details see Supplementary methods and Table S2). Our analysis combined these new DNA sequence data with all published Niphargus sequence data for 28S, COI and 16S available on GenBank on 1 June 2012 © 2014 John Wiley & Sons Ltd

A N C I E N T G R O U N D W A T E R F A U N A 1155 Fig. 1 Distribution of sampling locations from this study and published data included in the analysis.

(Lef
ebure et al. 2006, 2007; Fiser et al. 2008; H€anfling et al. 2008; Trontelj et al. 2009; Flot 2010; Flot et al. 2010; Hartke et al. 2011). Also included were published sequence data of the mitochondrial 12s rRNA region (12S) and the large subunit 18s rRNA (18S) for the taxa covered in the combined data set. In total, data were included from 78 described species, and several putative cryptic species, from 170 locations across the genus’ European range (Fig. 1, Tables S1 and S3, Supporting information). This included eight of the nine species (the ninth, N. boulangei, was too rare) that occur within 200 km of the British coast line (Table S3, Supporting information). The combined data set provided phylogeographical information (more than 10 populations) for eight of the 78 described species (Table S2, Supporting information; N. aquilex, N. fontanus, N. glenniei, N. kochianus, N. irlandicus, N. rhenorhodanensis, N. virei and N. schellenbergi). A further 11 species were covered by more than one specimen from 1 to 3 locations. A total of 36 taxa from nine amphipod families were used as out-groups to root the Niphargus phylogeny and provide calibration points for a molecular dating analysis (see Table S4, Supporting information). The out-group taxa include previously identified sister groups to Niphargidae (Englisch et al. 2003; Fiser et al. 2008) and © 2014 John Wiley & Sons Ltd

representatives of clades from a dated phylogeny of gammarid amphipods (Hou et al. 2011). We used the genes 28S, COI, 18S and elongation factor 1 alpha (EF-1a). The alignment of COI and EF-1a sequences was carried out using MUSCLE (Edgar 2004) in combination with MEGA version 5.05 (Tamura et al. 2011). Ribosomal genes were aligned with the software MAFFT version 6 (Katoh et al. 2002) using the alignment strategies Q-INS-i or E-INS-i.

Delineation of OTUs, multilocus alignments and phylogenies Cryptic diversity and taxonomic misclassification are common in Niphargus. We therefore used a DNA barcoding approach based on the two genes with the largest coverage (COI and 28S) to identify cryptic lineages within species and to delineate operational taxonomic units (OTUs) with independent evolutionary histories (for details see Supplementary materials). Many of these OTUs are likely to fulfil the criteria for separate species depending on the definition applied, but a discussion of species status is beyond the scope of this study. A multilocus alignment was created using representatives of OTUs of Niphargus and selected out-groups. One representative of each OTU was chosen at random for

1156 C . E . M C I N E R N E Y E T A L . inclusion in the supermatrix (Table S1, Supporting information). Amphipod out-groups included three representatives selected for each of the four Gammarus freshwater clades, six representatives of the marine Gammarus group, three representatives of the Baikalian gammarids and all out-groups used in Hou et al. (2011), providing 13 time-calibrated nodes. For each gene, all sequences of the selected taxa were aligned. Phylogenetic analysis of the multigene matrix was carried out using Bayesian analysis as implemented in MRBAYES V3.2 (Ronquist et al. 2012). Genes were used as partitions, and model parameters between partitions were unlinked. Two independent Markov chain Monte Carlo (MCMC) chains were run for 10 000 000 iterations each, sampling every 1000 iterations. The first 25% of each run was discarded as burn-in with the remaining samples pooled and used to create a maximum clade credibility tree.

This cryptic diversity is reflected in the OTUs used for the phylogenetic analysis. We therefore used the geographical coordinate of the individual chosen at random for the phylogenetic analysis as a proxy for the taxon’s geographical location. We estimated a phylogenetic model of evolution for the Niphargus in-group species where longitude and latitude were correlated using the computer program BayesTraits (Pagel et al. 2004). We ran the MCMC chain for one million iterations after apparent convergence sampling every 1000 iterations from the chain and repeated the analysis multiple times. We also simultaneously estimated the phylogenetic signal parameter k (Pagel 1999). The parameter k varies between 0 and 1, where 1 is interpreted as having the traits covary and zero means that the traits evolve independently of the phylogenetic relationships between species. Repeated analyses produced almost identical results; thus, we provide results from a single chain only.

Molecular dating using a Bayesian analysis

Geographical variation in species diversity and diversification rates

Bayesian Evolutionary Analysis Sampling Trees (BEAST) version 1.7.4 (Drummond et al. 2012) was used to generate an ultrametric phylogeny and estimate the time of the most recent common ancestor (TMRCA) for each node using a Bayesian MCMC analysis. Tree topology was constrained to that obtained from the MrBayes phylogenetic analysis. Genes were used as partitions, and substitution rates and clocks were unlinked in the analysis. An uncorrelated log-normal relaxed clock (Drummond et al. 2006) and a Yule speciation prior were used. A time-calibrated phylogeny of the amphipod group Gammaridae (Hou et al. 2011) was used to provide 11 external calibration points (for details see Supplementary materials).

Ancestral longitude and latitude reconstructions We used the Bayesian MCMC phylogenetic ancestral state reconstruction method introduced by Organ et al. (2007) to infer the geographical location of the MRCA for each node. The method was chosen because of its superior performance with phylogenetic trees that span millions of years (Montgomery et al. 2010). Similar methods have been used to infer ancestral longitudes and latitudes in a phylogenetic context (Lemey et al. 2009; Bouckaert et al. 2012). With exact geographical ranges mostly unknown, it was not possible to calculate range centroids. The range size of most Niphargus is small, however, usually 0.5 of nodes above the clade level are shown above branches. See Fig. 3 for PP of nodes within important clades. British and Irish taxa are marked with a red circle, and branches leading to them are highlighted in red; number in brackets refers to clade numbers in Fig. 3(b) and geographical distribution of major phylogenetic lineages; the exact location of the N. liasi sample is not known, but the species occurs in France (c).

The geographical distribution of MRCAs for nodes of different ages identified central France in north-western Europe as the origin of the Niphargus genus in the late Cretaceous (87 Ma). From there, the ancestral locations move with decreasing node age towards the south-east (Figs. 4b, S5 and Table S6, Supplementary information)).

Geographical variation in species diversity and diversification rates Investigation of the geographical variation in species diversity revealed that the number of Niphargus species varies greatly across different geographical areas from

one species in Spain to 136 in the Balkans (Table S5, Supporting information). In contrast to phylogenetic diversity, the species richness of the western region is significantly lower than that of the eastern region (P < 0.05). Investigation of the geographical variation in diversification rates shows that the number of nodes along each root-to-tip path in the Niphargus species-level phylogeny correlates significantly with distance towards the south-east [correlation coefficient (SD) = 0.18 (0.014), log Bayes factor = 9.8]. A log Bayes factor value of between 6 and 10 provides strong support for the hypothesis tested. Net-diversification rate in Niphargus therefore increases in a south-easterly direction. © 2014 John Wiley & Sons Ltd

A N C I E N T G R O U N D W A T E R F A U N A 1159 (a)

(b)

N aquilex A1 N aquilex A2 N aquilex B N aquilex C N aquilex D N aquilex F N aquilex E N schellenbergi 2 N schellenbergi 1

N irlandicus N glennei 10

0

Ma

30

(c)

20

10

0

(d)

Clade B from [7] N krameri Balkan N cf. krameri 1 N karamani N fontanus A1 N fontanus A2 N fontanus B N fontanus C

N kochianus A N kochianus B N kochianus C N kochianus D

10

0

Ma

Ma 30

20

10

0

Ma

Fig. 3 Geographical distribution of British and Irish operational taxonomic units (OTUs) and European sister taxa. The green and pink lines represent the maximum extent of the glacial ice sheets during the Devensian and Anglian glacial periods, respectively. Small black dots are sites of known distribution for each group; coloured dots represent sampled populations for each OTU. Partial ultrametric phylogenies from the BEAST analysis for each species complex are shown above maps for Niphargus irlandicus/N. glenniei group (a); N. aquilex/N. schellenbergi group (b); N. kochianus (c), N. fontanus (d).

Discussion Phylogenetic evidence for long-term persistence of Niphargus in NW Europe Palaeontological and genetic evidence suggests that the majority of surface fauna that currently live in Britain and Ireland originated from late Pleistocene/Holocene © 2014 John Wiley & Sons Ltd

dispersal from continental Europe (Wheeler 1977; Yalden 1982; Hewitt 2004). Endemic fauna are therefore rare (Pimm et al. 1995) and are restricted to a few surface invertebrate and vertebrate subspecies (e.g. the Irish hare, Reid 2011; the shelly freshwater whitefish, Kottelat & Freyhof 2007; and the avian Scottish crossbill, which is sometimes considered a species, Summers et al. 2007; see Table S7, Supporting information for

1160 C . E . M C I N E R N E Y E T A L . Table 1 (A) Estimates of the time of the most recent common ancestor (TMRCA) between British and Irish Niphargus taxa and their closest relatives based on a BEAST analysis. The prior used and mean and median estimates in millions of years (Ma) are given, including the upper and lower bounds of the highest posterior density (HPD) intervals. (B) Details of the external calibration points estimated from a subset of representative data from Hou et al. (2011) are provided Prior (A) Node N. aquilex E/F Tree prior N. aquilex (A1, A2)/B Tree prior N. aquilex A1/A2 Tree prior N. fontanus A1/A2 Tree prior N. irlandicus/N. glenniei Tree prior N. kochianus A/(B, C) Tree prior Niphargus root Uniform [45–558] (B) External calibration points from Hou et al. (2011) Node 1 Normal [5.0; 1] Node 2 Normal [30.0; 1] Node 3 Normal [44.9; 8] Node a Normal [80.3; 15] Node b Normal [61.3; 9] Node c Normal [42.8; 6] Node d Normal [36; 6] Node f Normal [25.6; 5] Node g Normal [28.2; 5] Node h Normal [32.9; 5] Node i Normal [36.6; 5]

Mean (Ma)

Median (Ma)

95% HPD lower

95% HPD upper

6.69 5.93 1.06 0.89 21.05 3.06 88.16

6.22 5.57 0.95 0.77 19.48 2.89 87.14

2.15 2.02 0.23 0.17 6.74 1.26 65.38

12.32 10.35 2.12 1.90 38.09 5.21 113.94

5.89 29.75 47.62 78.52 59.34 44.85 32.24 21.67 23.16 33.17 32.61

5.87 29.76 47.49 78.19 59.22 44.78 32.28 21.79 23.15 33.30 32.71

4.26 27.84 35.83 64.87 51.20 39.83 26.50 14.14 16.91 25.29 26.48

7.53 31.70 59.31 93.26 67.86 50.22 37.89 28.93 29.86 40.41 38.71

more examples). Critically, these fauna have only been present for a few tens of thousands of years. In contrast, our data indicate that groundwater contains by far the oldest endemic fauna, which have persisted for millions of years and represent a significant contribution to biodiversity. Furthermore, this ancient groundwater fauna has survived the extreme geological and climate changes that have occurred over the past 20 million years. Groundwater temperatures are influenced by air temperature (Figura et al. 2011) and can range from 0 to 6 °C in glacial and periglacial climates (Parsons 1970; Williams 1970) to >25 °C in areas with warm climates (Weyhenmeyer et al. 2000; Eberhard et al. 2009). Niphargids must therefore have survived a wide range of groundwater temperature conditions as climate changed between glacial and warm conditions. However,

temperature and chemistry change much more slowly in groundwater than in surface waters, and hence, groundwaters are buffered from temperature extremes and rapid hydrological and biological change (MacDonald et al. 2012). The relative stability of the subsurface environment may explain the persistence of groundwater invertebrates through changing climates. N. glenniei and N. irlandicus persisted in NW Europe throughout the Miocene surviving both glacial and extreme wet periods (Zachos et al. 2001a,b), which were associated with range shifts and local extinctions in other fauna (Zachos et al. 2001a, 2008; B€ ohme 2003; Klaus & Gross 2010). Together with N. aquilex B, they also persisted in Britain throughout the Pliocene when temperatures and sea levels were higher than today (Dwyer & Chandler 2009), and groundwaters would have been substantially warmer than they are now.

Fig. 4 Time-calibrated phylogeny of Niphargus generated with BEAST (out-group not shown), black dots indicate nodes with a posterior probability (PP) >0.5; clade numbers refer to clade numbers in Fig. 2a; British and Irish operational taxonomic units (OTUs) are marked in red (a); geographical location of the common ancestor for each node with a PP > 0.5 based on Bayesian-model-based ancestral state reconstruction; circle sizes are proportional to the age of nodes (b); schematic maps depicting some of the major palaeogeographical changes that occurred in Europe between 100 and 25 Ma; modified from Ron Blakey, NAU Geology (http://jan.ucc. nau.edu/rcb7/): 100 Ma, circle indicates putative location of Niphargus ancestor (c); 75 Ma isolation of Niphargus on a central European island and within the Tethys Sea; the question mark indicates the possibility that the Niphargus glenniei/N. irlandicus lineage became first isolated during this time on a north-western European island (d); 50 Ma spread of Niphargus across central Europe (e); 25 Ma spread of Niphargus to the Balkan and Italian Penisulas, circle indicates the location of the common ancestor of N. irlandicus and N. glenniei (f). © 2014 John Wiley & Sons Ltd

A N C I E N T G R O U N D W A T E R F A U N A 1161 (a) 1 2-5

(c) Origin of Niphargus...

6 7

8

9 10 11

Late Cretaceous (100 Ma) (d) ....followed by isolation

?

12

Late Cretaceous (75 Ma)

(b)

EOCENE

PALEOCENE

75

50

OLIGOCENE

MIOCENE

25

PLEI

LATE CRETACEOUS

PLIO

(e) Expansion across Central Europe

0 Mya

Early Eocene (50 Ma) (f) Origin of N glennei/ N irlandicus

Pleistocene (0.77-2.60 Mya) Pliocene (2.60-5.00 Mya)

South-east expansion of Superclade

Miocene (5.00-23.00 Mya) Oligocene (23.00-34.00 Mya) Eocene (34.00-56.00 Mya) 56.60 Mya (Paleocene) 57.70 Mya (Paleocene) 72.60 Mya (Cretaceous) 87.10 Mya (Cretaceous)

© 2014 John Wiley & Sons Ltd

0

195 390

780 Kilometers

Late Oligocene (25 Ma)

1162 C . E . M C I N E R N E Y E T A L . All the Niphargus lineages in Britain and Ireland have persisted throughout the multiple glaciations of the Quaternary. Our findings are congruent with those of Kornobis et al. (2010) who presented molecular evidence showing that the endemic subterranean amphipod Crangonyx islandicus has been present in Iceland for around 5 million years, surviving repeated glaciations. On the basis of the molecular analysis and the species distribution, Kornobis et al. (2010) suggest that C. islandicus may have survived in geothermally heated groundwaters associated with volcanic fissures. Our data demonstrate that some Niphargus populations have been resilient to climate changes that occur above ground in a region that is much less geothermally active. This suggests that groundwater ecosystems in general may have mechanisms that reduce the impacts of surface climate change, but our current understanding of these mechanisms is limited. During glaciations, groundwater taxa may have survived in caves or aquifers that were actively recharged by warm-based glaciers or proglacial rivers. Groundwater recharge from glaciers is well documented (Hutchinson & Thomas-Betts 1990; Boulton et al. 1995) and provides a source of oxygen and nutrients. However, these groundwaters would have been cooler than today, and therefore, surviving species must be resilient to these long-term variations in groundwater temperatures. Geothermal heating may have maintained some groundwaters at higher temperatures during glacial periods. For example, it has been suggested that areas of south-west England remained permafrost free in the last glaciation due to a high heat flux (Hutchinson & Thomas-Betts 1990) and there are small geothermal heat anomalies (approximately 2–6 °C) within 100 m of the surface in southern and eastern England (Busby et al. 2011). However, there is little relation between modern-day distributions of Niphargus in the British Isles and geothermally heated waters. For example, Niphargus are not recorded in Derbyshire in northern England where there are extensive geothermal springs and suitable geological habitats for invertebrates, and are present in areas of southern England where there is no evidence of geothermal warming of groundwater. Given the poor dispersal capabilities of Niphargus, it therefore seems unlikely that geothermal heating of groundwater was the only factor enabling their survival during glacial periods. A geothermal gradient of about 1 °C per 20–40 m (Anderson 2005) results in warmer waters at depth, which may have provided some protection against cold groundwaters if Niphargids were able to migrate to warmer, deeper waters. However, permeability and fracturing generally decrease substantially with depth (Williams et al. 2006; Jiang et al. 2010), resulting in limited groundwater circulation and low oxygen,

and therefore, the deep groundwater environment (>100 m) may not always provide a suitable habitat for invertebrates. Overall, it seems probable that surviving Niphargus species have some mechanism of adapting to changing groundwater temperatures. Modern-day occurrence of groundwater crustacea in subglacial refugia has been documented in Castleguard Cave, Canada, ca. 500 km north of the glacial limit, where groundwater temperatures are around 2 °C (Holsinger et al. 1983) and in lava caves beneath ice in Iceland (Kornobis et al. 2010). Nevertheless, other evidence indicates that the Pleistocene had a considerable negative impact on the distribution and survival of Niphargus. The British Niphargus species (Fig. 3) and N. virei in France (Foulquier et al. 2008) are largely found to the south of the maximum extent of the Anglian and Devensian glaciers, and species diversity in north-western Europe is relatively low, suggesting that some populations were eradicated during glacial or periglacial conditions.

Geographical origin of Niphargus and spread during the Cenozoic The geographical distribution of MRCAs for nodes of different ages showed a second unexpected pattern (Fig. 4b). The origin of Niphargus is in north-western Europe with the MRCA of all Niphargus in what is now central France in the late Cretaceous (87 Ma), when Europe consisted of a number of islands (Hay et al. 1999; R€ ogl 1999). The genus therefore pre-dates (and must have survived) the Cretaceous–Palaeogene mass extinctions of 65 Ma, possibly facilitated by a subterranean lifestyle. The schematic maps in Fig. 4c–f depict some of the major palaeogeographical changes that occurred between 100 and 25 Ma, although there were smaller-scale fluctuations in sea level and uplift superimposed on these broad patterns (Jarvis et al. 2002; Voigt et al. 2006). The ancestor of Niphargus probably colonized a central island (Fig. 4c), which was subsequently further inundated by the Tethys Sea (Fig. 4d). From there, the ancestral locations move with decreasing node age towards the south-east. During the Eocene, the retreating Tethys Sea provided the opportunity for Niphargus to spread in emerging freshwater aquifers (Fig. 4e). This is consistent with palaeogeographical models but is contrary to a previous hypothesis, which suggested that the enhanced species diversity in the northern parts of the Balkan Peninsula indicated an origin in south-east Europe (Karaman & Ruffo 1986). Our phylogenetically controlled analysis of diversification rates shows an increase in diversification in a south-easterly direction, thereby providing an alternative explanation for the enhanced species diversity in the Balkans. The timing of this diversification (around © 2014 John Wiley & Sons Ltd

A N C I E N T G R O U N D W A T E R F A U N A 1163 25 Ma) coincides with the closing of the Tethys Sea that had previously separated the Balkans and central Europe (R€ ogl 1999; Hrbek & Meyer 2003) and provided an opportunity for further dispersal towards the south-east (Fig. 4f). Available niche space in the geomorphologically complex Balkans may have enabled the high diversification rate, a mechanism that has also been suggested to explain diversification in other fauna (Hrbek & Meyer 2003).

Conclusions This study reveals the presence of an ancient endemic groundwater fauna in Britain and Ireland, where endemism is otherwise rare. The unusually high levels of endemism in groundwater fauna in northern latitudes identified by the study highlight the need to recognize this unique ecosystem, and its ancient organisms’ contribution to our understanding of climatic and palaeogeographical controls on global biodiversity. The extent to which Niphargus may be resilient to recent anthropogenic perturbations of groundwater ecosystems is unknown. However, the small ranges of these taxa shown in this study and others (Holsinger et al. 1983; Foulquier et al. 2008), and their smaller clutch sizes, delayed maturity, slower growth and lower population numbers compared to epigean relatives (Gibert et al. 1994), suggest that despite their ancient resilience, the European Niphargus fauna could now be vulnerable. Conservation policy measures to protect groundwater ecosystems in Europe lag far behind countries such as Australia. N. glenniei has been designated as a UK Biodiversity Action Plan (BAP) species, but other Niphargus species have no such recognition, and current European groundwater monitoring programmes do not consider groundwater ecosystems. The study also reveals an unusual biogeographical pattern within the Niphargus genus. The oldest and most phylogenetically diverse species occur in northern Europe where endemism is low in surface fauna, which are dominated by large-range species and post-glacial colonizers. In contrast, the species diversity is highest in southern Europe, indicating that once the genus dispersed to these areas, climatic and geomorphological conditions enabled a much higher diversification rate than has occurred in northern Europe. These groundwater organisms provide an unusual opportunity to improve our understanding of biological processes such as speciation, adaptation and convergence, and as narrow-range endemics, they allow further exploration of island biogeographical processes. Furthermore, our discovery that these groundwater species are the oldest known inhabitants of Britain and Ireland, persisting through millions of years of © 2014 John Wiley & Sons Ltd

changing climate may cast significant light on one of the major challenges facing the scientific community today, that of predicting the resilience of ecosystems to climate change (Chapin et al. 2000). Our findings show that groundwater fauna (or their habitats) are likely to have a highly variable response to the extinguishing effects of climate change. A more detailed knowledge of the mechanisms behind this variation could help us to understand the likely impacts of the current anthropogenically induced challenges to the biosphere.

Acknowledgements For financial support, we thank the EPA Ireland (STRIVE grant, project 2007-W-MS-1-S1), the Belgian Science Policy Department (BELSPO), the Linnean Society Systematics Association, the University of Hull, the British Cave Research Association, the British Geological Survey (this paper was published with permission of the Executive Director, British Geological Survey, NERC), Roehampton University and the Esm
ee Fairbairn Foundation. We thank Thierry Backeljau and Frank Fiers at the Royal Belgian Institute of Natural Sciences, and Bernie Doherty, Ian Anderson, Marlene Jahnke, Isabel Douterelo-Soler, G. Michel and A.J. Matthijs for technical support. At BGS we thank Andrew Newell for geological advice, Andrew McKenzie for assisting with geographical analyses and Debbie Allen and James Sorensen for sampling. We thank Tim Johns (Environment Agency) and MarieJo Dole-Olivier (Lyon University) for sampling and Joanna Baker for assistance with the ancestral state reconstruction. We thank Tim Guilford, Mark Pagel, David Lunt, Cock van Oosterhout and two anonymous referees for useful comments on the manuscript.

References Anderson MP (2005) Heat as a ground water tracer. Groundwater, 43, 951–968. B€ ohme M (2003) The Miocene Climatic Optimum: evidence from ectothermic vertebrates of Central Europe. Palaeogeography Palaeoclimatology Palaeoecology, 195, 389–401. B€ ohme M, Ilg A, Winklhofer M (2008) Late Miocene “washhouse” climate in Europe. Earth and Planetary Science Letters, 275, 393–401. Bouckaert R, Lemey P, Dunn M et al. (2012) Mapping the origins and expansion of the Indo-European language family. Science, 337, 957–960. Boulton GS, Caban PE, Vangijssel K (1995) Groundwater-flow beneath ice sheets. 1. Large-scale patterns. Quaternary Science Reviews, 14, 545–562. Busby J, Kingdon A, Williams J (2011) The measured shallow temperature field in Britain. Quarterly Journal of Engineering Geology and Hydrogeology, 44, 373–387. Chapin FS, Zavaleta ES, Eviner VT et al. (2000) Consequences of changing biodiversity. Nature, 405, 234–242. Cope JCW (1997) The Mesozoic and Tertiary history of the Irish Sea. In: Petroleum Geology of the Irish Sea and Adjacent Areas (eds Meadows NS, Trueblood SP, Hardman M, Cowan G), pp. 47–59. Geological Society of London, London.

1164 C . E . M C I N E R N E Y E T A L . Dole-Olivier M-J, Castellarini JF, Coineau N et al. (2009) Towards an optimal sampling strategy to assess groundwater biodiversity: comparison across six European regions. Freshwater Biology, 54, 777–796. Drummond AJ, Ho SYW, Phillips MJ, Rambaut A (2006) Relaxed phylogenetics and dating with confidence. Plos Biology, 4, 699–710. Drummond AJ, Suchard MA, Xie D, Rambaut A (2012) Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution, 29, 1969–1973. Dubilier N, Bergin C, Lott C (2008) Symbiotic diversity in marine animals: the art of harnessing chemosynthesis. Nature Reviews Microbiology, 6, 725–740. Dwyer GS, Chandler MA (2009) Mid-Pliocene sea level and continental ice volume based on coupled benthic Mg/Ca palaeotemperatures and oxygen isotopes. Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences, 367, 157–168. Eberhard SM, Halse SA, Williams MR, Scanlon MD, Cocking J, Barron HJ (2009) Exploring the relationship between sampling efficiency and short-range endemism for groundwater fauna in the Pilbara region, Western Australia. Freshwater Biology, 54, 885–901. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research, 32, 1792–1797. Englisch U, Coleman CO, W€agele JW (2003) First observations on the phylogeny of the families Gammaridae, Crangonyctidae, Melitidae, Niphargidae, Megaluropidae and Oedicerotidae (Amphipoda, Crustacea), using small subunit rDNA gene sequences. Journal of Natural History, 37, 2461–2486. Figura S, Livingstone DM, Hoehn E, Kipfer R (2011) Regime shift in groundwater temperature triggered by Arctic Oscillation. Geophysical Research Letters, 38, L23401. Fiser C, Sket B, Trontelj P (2008) A phylogenetic perspective on 160 years of troubled taxonomy of Niphargus (Crustacea: Amphipoda). Zoologica Scripta, 37, 665–680. Flot J (2010) Toward a molecular taxonomy of the amphipod genus Niphargus: examples of use of DNA sequences for species identification. Bulletin de la Soci
et
e des sciences naturelles de l’Ouest de la France, 32, 62–68. Flot JF, Worheide G, Dattagupta S (2010) Unsuspected diversity of Niphargus amphipods in the chemoautotrophic cave ecosystem of Frasassi, central Italy. Bmc Evolutionary Biology, 10, 171. Foulquier A, Malard F, Lefebure T, Gibert J, Douady CJ (2008) The imprint of Quaternary glaciers on the present-day distribution of the obligate groundwater amphipod Niphargus virei (Niphargidae). Journal of Biogeography, 35, 552–564. Freckleton RP, Phillimore AB, Pagel M (2008) Relating traits to diversification: a simple test. American Naturalist, 172, 102–115. Galassi DMP, Stoch F, Fiasca B, Di Lorenzo T, Gattone E (2009) Groundwater biodiversity patterns in the Lessinian Massif of northern Italy. Freshwater Biology, 54, 830–847. Gibert J, Stanford JA, Dole-Olivier M-J, Ward JV (1994) Basic attributes of groundwater ecosystems and prospects for research. In: Groundwater Ecology (eds Gibert J, Danielopol DL & Standford JA), pp. 8–42. Academic Press, San Diego, California.

Gupta S, Collier JS, Palmer-Felgate A, Potter G (2007) Catastrophic flooding origin of shelf valley systems in the English Channel. Nature, 448, U342–345. H€ anfling B, Douterelo-Soler I, Knight L, Proudlove G (2008) Molecular studies on the Niphargus kochianus group (Crustacea: Amphipoda: Niphargidae) in Great Britain and Ireland. Cave and Karst Science, 35, 35–40. Hartke TR, Fiser C, Hohagen J, Klebera S, Hartmannd R, Koenemanne S (2011) Morphological and molecular analyses of closely related species in the stygobiontic genus Niphargus (Amphipoda). Journal of Crustacean Biology, 31, 701–709. Hay WH, DeConto RM, Wold CN, et al. (1999) Alternative global Cretaceous paleogeography. In: Evolution of the Cretaceous Ocean-Climate System Pages (eds Barrera E & Johnson CC), pp. 1–47. Geological Society of America, Boulder, Colorado. Hewitt GM (2004) Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 359, 183–195. Hof C, Br€ andle M, Brandl R (2008) Latitudinal variation of diversity in European freshwater animals is not concordant across habitat types. Global Ecology and Biogeography, 17, 539–546. Holsinger JR, Mort JS, Recklies AD (1983) The subterranean crustacean fauna of Castleguard Cave, Columbia icefields, Alberta, Canada, and its zoogeographic significance. Arctic and Alpine Research, 15, 543–549. Hou ZE, Sket B, Fiser C, Li S (2011) Eocene habitat shift from saline to freshwater promoted Tethyan amphipod diversification. Proceedings of the National Academy of Sciences of the United States of America, 108, 14533–14538. Hrbek T, Meyer A (2003) Closing of the Tethys Sea and the phylogeny of Eurasian killifishes (Cyprinodontiformes: Cyprinodontidae). Journal of Evolutionary Biology, 16, 17–36. Hutchinson JN, Thomas-Betts A (1990) Extent of permafrost in southern Britain in relation to geothermal flux. Quarterly Journal of Engineering Geology, 23, 387–390. Illies J (1978) Limnofauna Europaea, 2nd edn. Gustav Fischer Verlag, Stuttgart. Jarvis I, Mabrouk A, Moody RTJ, de Cabrera S (2002) Late Cretaceous (Campanian) carbon isotope events, sea-level change and correlation of the Tethyan and Boreal realms. Palaeogeography Palaeoclimatology Palaeoecology, 188, 215–248. Jiang X-W, Wang X-S, Wan L (2010) Semi-empirical equations for the systematic decrease in permeability with depth in porous and fractured media. Hydrogeology Journal, 18, 839–850. Karaman GS, Ruffo S (1986) Amphipoda: Niphargus group (Niphargidae sensu Bousfield, 1982). In: Stygofauna Mundi (ed. Botosaneanu L), pp. 514–534. E.J. Brill, Leiden. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Research, 30, 3059–3066. Klaus S, Gross M (2010) Synopsis of the fossil freshwater crabs of Europe (Brachyura: Potamoidea: Potamidae). Neues Jahrbuch f€ ur Geologie und Pal€aontologie-Abhandlungen, 256, 39–59. Kornobis E, P
alsson S, Kristj
ansson BK, Svavarsson J (2010) Molecular evidence of the survival of subterranean amphipods (Arthropoda) during ice age underneath glaciers in Iceland. Molecular Ecology, 19, 2516–2530.

© 2014 John Wiley & Sons Ltd

A N C I E N T G R O U N D W A T E R F A U N A 1165 Kottelat M, Freyhof J (2007) Handbook of European Freshwater Fishes. Publications Kottelat, Cornol, Switzerland. Lef
ebure T, Douady CJ, Gouy M, Trontelj P, Briolay J, Gibert J (2006) Phylogeography of a subterranean amphipod reveals cryptic diversity and dynamic evolution in extreme environments. Molecular Ecology, 15, 1797–1806. Lef
ebure T, Douady CJ, Malard F, Gibert J (2007) Testing dispersal and cryptic diversity in a widely distributed groundwater amphipod (Niphargus rhenorhodanensis). Molecular Phylogenetics and Evolution, 42, 676–686. Lemey P, Rambaut A, Drummond AJ, Suchard MA (2009) Bayesian phylogeography finds its roots. Plos Computational Biology, 5, e1000520. L
opez-Garcia P, Philippe H, Gail F, Moreira D (2003) Autochthonous eukaryotic diversity in hydrothermal sediment and experimental microcolonizers at the Mid-Atlantic Ridge. Proceedings of the National Academy of Sciences of the United States of America, 100, 697–702. Louwye S, Foubert A, Mertens K, Van Rooij D, The IODP Expedition 307 Scientific Party (2008) Integrated stratigraphy and palaeoecology of the lower and Middle Miocene of the Porcupine Basin. Geological Magazine, 145, 321–344.
Dochartaigh BE,
Taylor RG MacDonald AM, Bonsor HC, O (2012) Quantitative maps of groundwater resources in Africa. Environmental Research Letters, 7, 024009. Martin P, De Broyer C, Fiers F, Michel G, Sablon R, Wouters K (2009) Biodiversity of Belgian groundwater fauna in relation to environmental conditions. Freshwater Biology, 54, 814–829. Montgomery SH, Capellini I, Barton RA, Mundy NI (2010) Reconstructing the ups and downs of primate brain evolution: implications for adaptive hypotheses and Homo floresiensis. BMC Biology, 8, 9. Organ CL, Shedlock AM, Meade A, Pagel M, Edwards SV (2007) Origin of avian genome size and structure in non-avian dinosaurs. Nature, 446, 180–184. Pagel M (1999) Inferring the historical patterns of biological evolution. Nature, 401, 877–884. Pagel M, Meade A, Barker D (2004) Bayesian estimation of ancestral character states on phylogenies. Systematic Biology, 53, 673–684. Parsons ML (1970) Groundwater Thermal Regime in a glacial complex. Water Resources Research, 6, 1701–1720. Pimm SL, Russell GJ, Gittleman JL, Brooks TM (1995) The future of biodiversity. Science, 269, 347–350. Provan J, Bennett KD (2008) Phylogeographic insights into cryptic glacial refugia. Trends in Ecology & Evolution, 23, 564–571. Reid N (2011) European hare (Lepus europaeus) invasion ecology: implication for the conservation of the endemic Irish hare (Lepus timidus hibernicus). Biological Invasions, 13, 559–569. Robertson AL, Smith JWN, Johns T, Proudlove GS (2009) The distribution and diversity of stygobites in Great Britain: an analysis to inform groundwater management. Quarterly Journal of Engineering Geology and Hydrogeology, 42, 359–368. R€ ogl F (1999) Mediterranean and Paratethys. Facts and hypotheses of an Oligocene to Miocene paleogeography (short overview). Geologica Carpathica, 50, 339–349. Ronquist F, Teslenko M, van der Mark P et al. (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference and model

© 2014 John Wiley & Sons Ltd

choice across a large model space. Systematic Biology, 61, 539–542. Summers RW, Dawson RJG, Phillips RE (2007) Assortative mating and patterns of inheritance indicate that the three crossbill taxa in Scotland are species. Journal of Avian Biology, 38, 153–162. Tamura K, Peterson D, Peterson N, Nei M, Kumar S (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution, 28, 2731–2739. Trontelj P, Douady CJ, Fiser C et al. (2009) A molecular test for cryptic diversity in ground water: how large are the ranges of macro-stygobionts? Freshwater Biology, 54, 727–744. Trontelj P, Blejec A, Fiser C (2012) Ecomorphological convergence of cave communities. Evolution, 66, 3852–3865. V€ ain€ ol€ a R, Witt JDS, Grabowski M, Bradbury JH, Jazdzewski K, Sket B (2008) Global diversity of amphipods (Amphipoda; Crustacea) in freshwater. Hydrobiologia, 595, 241–255. Van Dover CL, German CR, Speer KG, Parson LM, Vrijenhoek RC (2002) Evolution and biogeography of deep-sea vent and seep invertebrates. Science, 295, 1253–1257. Voigt S, Gale AS, Voigt T (2006) Sea-level change, carbon cycling and palaeoclimate during the Late Cenomanian of northwest Europe; an integrated palaeoenvironmental analysis. Cretaceous Research, 27, 836–858. Weyhenmeyer CE, Burns SJ, Waber HN et al. (2000) Cool glacial temperatures and changes in moisture source recorded in Oman groundwaters. Science, 287, 842–845. Wheeler A (1977) The origin and distribution of the freshwater fishes of the British Isles. Journal of Biogeography, 4, 1–24. Williams JR (1970) Ground Water in the Permafrost Regions of Alaska, p. 83. U.S. Geological Survey, Professional Paper 696. Williams A, Bloomfield J, Griffiths K, Butler A (2006) Characterising the vertical variations in hydraulic conductivity within the Chalk aquifer. Journal of Hydrology, 330, 53–62. Yalden DW (1982) When did the mammal fauna of the British Isles arrive. Mammal Review, 12, 1–57. Zachos J, Pagani M, Sloan L, Thomas E, Billups K (2001a) Trends, rhythms, and aberrations in global climate 65 Ma to present. Science, 292, 686–693. Zachos JC, Shackleton NJ, Revenaugh JS, P€ alike H, Flower BP (2001b) Climate response to orbital forcing across the Oligocene-Miocene boundary. Science, 292, 274–278. Zachos JC, Dickens GR, Zeebe RE (2008) An early Cenozoic perspective on greenhouse warming and carbon-cycle dynamics. Nature, 451, 279–283.

B.H., G.S.P., A.L.R. and L.M. conceived the study; B.H., C.V., T.M. and A.L.R. conducted analyses; C.E.M., B.H. and S.M. carried out laboratory work and assembled the data; C.E.M., S.M., L.R.F.D.K., L.M., A.L.R., J.A., J.S.G.D. and K.E. collected the samples and contributed data; B.H., L.M., A.L.R., G.S.P. and C.E.M. wrote the paper; all authors commented on the final draft.

1166 C . E . M C I N E R N E Y E T A L .

Data accessibility DNA sequences (de novo sequencing): GenBank accessions KC315528; KC315548; KC315569; KC315574; KC315577; KC315583; KC315590; KC315598; KC315599; KC315602–KC315708; KC733825. Sampled Niphargus taxa; sampling locations; complete list of GenBank accession numbers including previously published data (Table S1, Supporting information): Dryad doi:10.5061/ dryad.82vg8. Out-group taxa including GenBank accession numbers: uploaded as online supplemental material. Final DNA supermatrix alignment used for phylogenetic analysis: Dryad doi:10.5061/dryad.82vg8.

Supporting information Additional supporting information may be found in the online version of this article. Table S1 Data included in the analyses including post-hoc taxonomic classification as operational taxonomic unit (OTU), subset of OTUs used in the supermatrix analysis, geographic information for samples including country, location and latitude and longitude coordinates decimalized are provided (collected from publications or approximated from sample site name and country information). Genes included (28S, COI, 12S, 16S) and GenBank accession numbers for sequences, and the phylogenetic lineage, and source of the published data used in this study. (Separate Excel spreadsheet) are listed. Table S2 List of PCR primer pairs and sequencing primers used in this study.

Table S3 Overview of genetic data from this study and the additional eight published data sets included in the analyses. Table S4 Outgroup taxa, genes used (28S, COI, 18S, EF-1a) and the Genbank accession numbers for sequences used in the phylogenetic analyses. Table S5 Total species richness of Niphargus and species richness standardised for area across eight geographic regions in Europe. Table S6 Geographical coordinates of the location of the MRCA for each node with a posterior probability (PP) >0.5 based on the Bayesian ancestral state reconstruction. Table S7 Examples of endemic taxa of Britain and Ireland, including taxonomic group and details of the distribution. Fig. S1 Phylogeny of cytochrome oxidase subunit I haplotypes inferred using the Maximum Composite Likelihood Method in combination with Neighbour-Joining. Fig. S2 Phylogeny of 28S haplotypes inferred using the Maximum composite Likelihood method in combination with Neighbour-Joining. Fig. S3 Multi-gene phylogeny of the species level data set based on a MrBayes analysis. Fig. S4 Maximum clade credibility chronogram inferred from a molecular dating analysis using the BEAST software. Fig. S5 ID’s for the nodes used in the ancestral state reconstruction of latitude and longitude. Data S1 Supplementary methods.

© 2014 John Wiley & Sons Ltd